Tag
#benchmarks
2posts

AI & Agents10 min read
Google Harness² Explained: Recursive Agent Tuning With Frozen Models
Deep dive into Google Harness²: recursive two-level agent optimization (task & harness) with frozen model weights, boosting performance without ground truth.
#ai-agents#google-research#benchmarks

AI & Agents5 min read
GLM-5.2 just took the open-weights crown - and it thinks a lot
Z.ai's GLM-5.2 tops Artificial Analysis's open-weights leaderboard at 51 points, but HN is arguing about token bloat and whether benchmarks lie.
#glm-5-2#open-weights#llm